Cardiovascular risk emerges below the diagnostic threshold for diabetes: insights from the Western Denmark Heart Registry
K K W Olesen, P T Tonnesen, C Gyldenkerne, N Sattar, H T Sorensen, M MaengAbstract
Background
Diabetes is defined by glycated haemoglobin (HbA1c) thresholds based on the risk of microvascular complications. However, it remains unclear at which HbA1c level macrovascular risk increases among patients without known diabetes, particularly when comparing patients with and without coronary artery disease (CAD).
Purpose
To examine whether the HbA1c threshold associated with cardiovascular complications is lower than the conventional diagnostic cut-off for diabetes, and whether this differs according to CAD status.
Methods
We included patients without previously diagnosed diabetes who underwent coronary angiography in Western Denmark from 2011 to 2021. HbA1c was measured within six months prior to angiography. We examined 10-year rates of non-fatal myocardial infarction, coronary revascularization (percutaneous coronary intervention or coronary artery bypass grafting), cardiovascular death, and all-cause death according to HbA1c level. Hazard ratios were estimated using cause-specific Cox regression analysis adjusted for age, sex, hypertension, peripheral artery disease, previous ischemic stroke, smoking, and CAD. Associations between HbA1c and outcomes were modelled using restricted cubic splines with 38 mmol/mol (5.6%) as reference. Analyses were additionally stratified by CAD status.
Results
Among 55,570 patients, 39,957 (72%) had CAD, and 1,112 (2.0%) had undiagnosed diabetes [i.e. HbA1c ≥48 mmol/mol (6.5%)]. Median HbA1c was 38 mmol/mol (Q1-Q3: 35-40). Median follow-up was 5.1 years (Q1-Q3: 3.0-7.2). HbA1c showed a U-shaped association with fatal outcomes, namely cardiovascular death and all-cause mortality, while there were broadly more linear associations between HbA1c and the non-fatal outcomes of myocardial infarction and coronary revascularization (Figure 1). Cardiovascular risk increased at glycaemic levels below the diagnostic threshold for diabetes across all outcomes. Associations were strongest in patients with CAD, whereas patterns were similar but attenuated among those without CAD (Figure 2).
Conclusions
In keeping with prior data, HbA1c demonstrates U-shaped associations with fatal CV outcomes, in those with or without CAD, likely in part due to reverse causality. The association between HbA1c and non-fatal CV outcomes of MI and revascularisation were more linear and were apparent even below 48 mmol/mol, likely reflecting in part risk factors that cluster with rising HbA1c across and beyond the normal range. These findings indicate that macrovascular risk becomes apparent before overt diabetes is diagnosed and may support closer cardiovascular risk assessment and preventive strategies in patients with modestly elevated HbA1c.Figure 1Figure 2